1

Stop caring about consciousness

Peter Carruthers

Abstract: The best empirically-grounded theory of first-personal phenomenal consciousness is global

workspace theory. This, combined with the success of the phenomenal concept strategy, means that

consciousness can be fully reductively explained in terms of globally broadcast nonconceptual content.

So there are no qualia (and there is no mental paint). As a result, the question of which other creatures

besides ourselves are conscious is of no importance, and doesn’t admit of a factual answer in most

cases. What is real, and what does matter, is a multidimensional similarity pace of functionally organized

minds.

Keywords: access consciousness, animal minds, global workspace, phenomenal concept strategy,

phenomenal consciousness, qualia

1. Introduction

Our topic is phenomenal consciousness. This is the kind of consciousness that gives rise to “hard”-

problem thought experiments, such as the conceivability of zombies and the explanatory gap.

Phenomenal consciousness is basically first-personal. Talk of the “feel” of experience and of what

experiences are “like” are really just invitations to one’s hearers to pay attention to their own

experiences to verify what one is saying from their own introspection. It is these introspection-based

first-person thoughts that define our subject matter. Secondarily, we can also formulate a public

concept of phenomenal consciousness, of course. My view is that this is best understood as: whatever

property it is that gives rise to the “hard”-problem thought experiments (Carruthers & Veillet, 2017).

That is what I will assume here.

There are numerous other ways in which the term “consciousness” is used, of course, and these

need to be sharply distinguished from our target (Rosenthal, 2005). Indeed, the field of consciousness

studies is littered with confusion because people fail to keep sight of these important distinctions. First

of all, phenomenal consciousness is a variety of mental-state consciousness. It is the mental states of a

person that are like something to undergo, and that have a distinctive subjective feel. People are

phenomenally conscious derivatively, by virtue of undergoing phenomenally conscious mental states.

Mental states can also be conscious in another sense, however, by being accessible to inform reasoning

2

and decision making, as well as giving rise to long-term memories and issuing in verbal reports. This is

access consciousness, which is third-personal (functionally defined), and is at least conceptually distinct

from phenomenal consciousness. As we will see shortly, however, access consciousness, in its

contemporary guise of global workspace theory, provides the best candidate for explaining phenomenal

consciousness.

People (and other animals) can also be said to be conscious in a quite different sense. This is

creature consciousness, which admits of both transitive and intransitive varieties. Intransitive creature

consciousness is a matter of being awake rather than asleep, or conscious rather than comatose. Almost

all living creatures undergo sleep / awake cycles, of course, and so are sometimes conscious in this

sense. Transitive creature consciousness, on the other hand, coincides with perceptual consciousness. A

creature can be conscious of some aspect of the world or of its own body, meaning that it perceives it.

Thus a cat that sees a mouse scuttling towards its hole can be said to be conscious of the mouse. Both

forms of creature consciousness, like access consciousness, are third-personal and functionally defined.

And likewise both (also like access consciousness) admit of degrees. A creature can be more-or-less

awake (one can be partially conscious), and perceptual consciousness can be more or less rich in

content, both within an individual and across different species of animal.1

There is much to learn about perception in ourselves and other creatures, of course, as well as

about the varieties and richness of perceptual consciousness across species. Likewise, there is much yet

to be learned about the nature and function of sleep and dreaming, as well as about the mechanisms

that govern waking and alertness. These may be hard scientific questions, well worth studying. But they

raise no problems for a scientific-naturalist, basically physicalist, world view, any more than mentality

itself does. Third-personal forms of consciousness are straightforwardly amenable to scientific study and

scientific explanation.

Phenomenal consciousness, in contrast, is thought by many to give rise to deep problems for

physicalist forms of naturalism. This is because one can see that no matter how much one learns about

the physical, functional, and representational properties of one’s own phenomenally conscious states,

one will still experience an explanatory gap between the former and the latter. For one will be able to

1 In Carruthers (2019) I place quite a bit of weight on the fact that all of the functionally-defined forms of

consciousness admit of degrees across species, whereas first-person phenomenal consciousness is all-or-nothing—

a given mental state in oneself is either like something to undergo or it isn’t. But I won’t need to rely on this point

in the discussion that follows.

3

think, “all that [physical / functional / representational] stuff might have been true, and yet this feel

might have been different or absent.” Likewise, there seems to be no difficulty in conceiving of a

creature that is like oneself in all physical, functional, and representational respects but who

nevertheless lacks these kinds of (phenomenally conscious) mental states altogether. That is, one can

conceive of a zombie version of oneself. In consequence, many have wanted to conclude that

phenomenal consciousness involves a distinctive set of nonphysical properties (qualia) that are

responsible for the felt qualities of one’s experience (Chalmers, 1996).

I will argue here, however, that the so-called “hard” problem of consciousness results entirely

(and, in a sense, trivially) from the differing perspectives one can take on one-and-the-same experiential

state (third person versus first person). No extra properties enter the world with phenomenal

consciousness, and nothing mysterious is implicated. Moreover, when one attempts to project one’s

first-person conception of phenomenal consciousness across species, the result is indeterminacy, and

there are no real facts of the matter waiting to be discovered.

2. Global workspace theory

Over the past 40 years, numerous labs around the world have joined the hunt for the neural correlates

of (phenomenal) consciousness, using a variety of scientific methods and technologies. The result has

been a proliferation of theories, including global workspace theory (Baars, 1988; Dehaene, 2014),

integrated information theory (Tononi, 2008; Tononi & Koch, 2015), higher-order thought theory

(Rosenthal, 2005; Lau & Rosenthal, 2011), and a theory that locates phenomenal consciousness in the

contents of a form of fragile short-term memory (Block, 2007, 2011). I have examined these (and other)

theories in some detail elsewhere (Carruthers, 2019). Here I will just sketch some of the evidence that

supports global workspace theory, as well as some of the evidence against Block’s suggestion that

consciousness coincides with of a richer form of short-term memory.

Especially important in the investigation of the neural correlates of consciousness have been

experimental paradigms in which the stimuli can be kept exactly the same, but where consciousness can

be either present or absent. For example, in backward masking experiments, very briefly presented

stimuli are followed immediately by a follow-up stimulus, which often has the effect of rendering the

first image invisible. If the timing and intensity of the stimuli are carefully titrated, one can achieve a

situation in which one-and-the-same type of stimulus is consciously perceived on about 50 per cent of

trials. One can then use fMRI, MEG, and/or EEG to investigate what happens differently in the brain

4

when the same stimuli are experienced consciously as opposed to unconsciously.2

The general finding in such experiments is that in cases where a stimulus remains unconscious

(albeit being successfully processed up the level of semantic meaning, perhaps, as well as priming

subsequent responses), there is reverberating activity in mid- and high-level sensory cortices at the

“back” of the brain that rapidly fades away. When the stimulus is consciously experienced, in contrast,

there is coordinated activity involving prefrontal, parietal, and temporal cortices as well. Multiple forms

of inquiry suggest that attention plays an important role in this process. Attentional signals boost the

activity of target populations of neurons while suppressing competitors, generally resulting in a sort

global ignition that makes the targeted contents widely available to a range of different systems capable

of remembering, describing, or otherwise responding to them. In consequence, many theorists now

identify consciousness with the contents of working memory (or the contents of the “global workspace”;

Dehaene, 2014; Carruthers, 2019; Mashour et al., 2020).3

Critics have objected that the prefrontal activity one observes when stimuli are consciously

experienced might not be a correlate of consciousness itself, but rather the engagement of sustained

attention and/or response preparation (Michel, 2017). A variety of clever experiments make this

interpretation unlikely, however. One of these required participants to make a response on every trial,

whether the stimulus was consciously experienced or not (Salti et al., 2015). Since participants often

responded correctly at above-chance levels even when the stimuli were unseen, one can assume that a

blindsight-like action-facilitation phenomenon was at work. The experimenters then trained pattern

classifiers on the resulting brain-imaging data to see where in the brain the content of the stimulus was

represented when consciously seen, contrasting this with cases where participants responded correctly

but without undergoing conscious experience. In the former case, those contents reached much deeper

into prefrontal regions while being processed more robustly there.

2 fMRI (functional magnetic resonance imaging) has quite good spatial but poor temporal resolution. EEG

(electroencephalogram) has excellent temporal resolution but poor spatial resolution. MEG

(magnetoencephalography) has intermediate virtues and drawbacks, both spatial and temporal. Sometimes more

than one of these methods can be combined together in a single experiment. 3 Or at least, they identify consciousness with actively-maintained forms of working memory. For there is recent

evidence that information can be passively retained, and readily recovered, for short periods of time in the form of

temporarily-altered synaptic weights (Lewis-Peacock et al., 2012; Trübutschek et al., 2018). I myself think that this

shouldn’t really be counted as a form of working memory at all. Rather, it belongs alongside the phenomena that

motivated people to talk about “long-term working memory” (Ericsson & Kintsch, 1995).

5

One particular style of experiment is especially telling. This makes use of binocular rivalry, and

has been conducted with monkeys as well as humans. In binocular rivalry, distinct images are

simultaneously presented to the two eyes. At any one time, only one of the images is consciously

experienced, with experience flipping from one image to another at irregular intervals because of neural

adaptation. Given the right experimental conditions and/or as a result of pretesting, one can sometimes

reliably predict which stimulus will be experienced within a specific time frame. In two recent

experiments conducted with monkeys, the experimenters recorded from populations of neurons in

higher-level visual areas, and also in regions of lateral prefrontal cortex, after first discovering which

populations represented which of the two stimuli. The finding was that prefrontal content-specific

neurons are active in purely-passive viewing conditions when and only when the corresponding stimulus

is dominating (Panagiotaropoulos et al., 2012; Kapoor et al., 2020).4

Block (2007, 2011) concedes the evidence that access consciousness involves prefrontal

representation, while arguing that phenomenal consciousness is nevertheless something distinct:

accessible for report when targeted by attention, but with richer contents than can enter working

memory at any one time. For when a rich stimulus array of letters or shapes is briefly presented on a

screen, participants report conscious experience of the entire array, despite only being able to identify

about four items (the standard limit for the contents of working memory). Yet, if the location of any one

target is retrospectively cued, participants can report it, suggesting that their conscious experience was,

indeed, rich, overflowing what can simultaneously enter working memory (Sperling, 1960; Landman et

al., 2003; Sligte et al. 2008). Block thus identifies phenomenal consciousness with the rich contents of a

fragile short-term memory system located in mid- or higher-level sensory cortices.

Critics have acknowledged the existence of fragile short-term sensory memory, and its richness;

but they have denied that the data demonstrate that its contents are conscious. Rather, the rich

stimulus array can be represented unconsciously until a cue targets a particular item with attention,

resulting in it becoming conscious (Cohen & Dennett, 2011; Cohen et al., 2016). People’s impression of

richness when experiencing the initial stimulus can be explained by a combination of consciously-

4 One might worry that questions about phenomenal consciousness in animals are begged by appealing to

experiments conducted with monkeys. But in fact the experiments just presuppose a (limited) form of access

consciousness in monkeys, and their evidential role is to demonstrate that the prefrontal activity observed when

humans undergo phenomenally conscious states is not due to either response preparation or the engagement of

attention. For attention is engaged whichever of the two stimuli dominates in binocular rivalry; and the monkeys

remained entirely passive: no response was required of them.

6

experienced gist content (e.g. “a bunch of letters arranged in rows”) together with conscious experience

of just a few items (Cohen et al., 2011; Wu & Wolfe, 2018)—no doubt reinforced by the background

folk-psychological assumption of the richness of experience in general, resulting from ignorance of the

frequent saccades made by one’s eyes, resulting in constant sampling of the environment.

The data Block appeals to in his support are not persuasive, then. In addition, one reason for

positively rejecting Block’s claims is that other retro-cuing experiments have found that a cue that is

presented following stimulus offset can improve conscious experience of even a single very faint

stimulus (Sergent et al., 2013; Thibault et al., 2016; Xia et al., 2016). If the contents of fragile short-term

memory are always conscious, then a retro-cue should be ineffective: the stimulus should already be

conscious. And if it is replied that the stimulus may give rise to a representation in short-term memory

that is too faint to be conscious in the absence of directed (cued) attention, then that undermines the

probative value of the original Sperling-style experiments. For it can be said that the contents of fragile

short-term memory, in general, remain unconscious until targeted by attention.

In my view the cleanest demonstration of the falsity of a view that identifies consciousness with

the relatively rich contents of fragile short-term memory is provided by Tsubomi et al. (2013).

(Unfortunately, the relevance of their data for our topic seems to have been missed by many.) They first

tested working memory limits for colored squares in a standard way, by presenting an array of them,

followed by a retention interval, followed by a single square selected from one of the original positions.

But now one half of the square retained the original color, whereas the other half was drawn from

elsewhere in the initial array. Participants just had to say which was the original color of that square.

Tsubomi and colleagues found the standard four-item limit overall. But then they re-ran the test without

any retention interval: at the very moment when the other squares disappeared, the remaining one

became color-divided, with one half retaining the original color and the other drawn from elsewhere in

the initial array. There was no difference in performance: it still levelled off when more than four items

overall were presented (and by just the same amount). This finding would be mysterious if it were true,

as Block claims, that an entire array of up to around eight or more items can be consciously experienced

at the same time. For if the item that remains when the others disappear were already consciously

experienced, then one would think one would immediately know which side of it was the original color

and which side had just changed.5

5 I have simplified my presentation of these findings for present purposes. In fact, participants kept their eyes on a

central fixation point, and the squares were presented on either side of it. Two squares then remained on the

7

In what follows, then, for the reasons briefly sketched here (together with others; see

Carruthers, 2019), I propose to take for granted that phenomenal consciousness coincides with the

contents of the global workspace (with access consciousness, in other words). Two further assumptions

can be put in place at this point, however. First, while I have not argued for this here, Carruthers &

Veillet (2017) show that only nonconceptual globally broadcast contents make a constitutive (rather

than a causal) contribution to the phenomenal properties of experience. I shall likewise make this

assumption in what follows, although nothing substantive will turn on it for our purposes.

Then second, I propose to assume that phenomenally conscious properties line up one-for-one

with globally broadcast representational contents. That is to say, there are no differences in

phenomenal consciousness that aren’t accompanied by representational differences; and likewise,

when representations figure within the global workspace, there are no representational differences in

nonconceptual content that don’t impact phenomenology. So there is no “mental paint” (Block, 2010),

and the phenomenal properties of things like moods and the painfulness of pain can be captured in

representational terms. This has been argued elsewhere by myself and others (Tye, 1995, 2000; Cutter &

Tye, 2011; Carruthers, 2018, 2019). I don’t propose to revisit those arguments here.

3. The phenomenal concept strategy

I propose, now, to take for granted that phenomenal consciousness correlates one-to-one with globally

broadcast nonconceptual content. From this it doesn’t follow, of course, that phenomenal

consciousness is globally broadcast nonconceptual content. But that is the view I will sketch arguments

for here. More strictly, I will suggest that all of the facts surrounding phenomenal consciousness

(including the “hard”-problem thought experiments) can be fully reductively explained in third-person

psychological terms, allowing us to conclude that there is at least a local identity between globally

broadcast nonconceptual content and phenomenal consciousness in humans.

The basic explanatory strategy, here, is long-standing and familiar (Loar, 1990; Tye, 1995, 2000;

Balog 1999, 2012; Carruthers, 2000; Papineau, 2002; Carruthers & Veillet, 2007; Prinz, 2012). It is to

argue that the so-called “hard” problems (the explanatory gap, the conceivability of zombies, and the

rest) result, not from any special properties attaching to our conscious experiences (that is, not from the

screen, one in each hemi-field. The finding of an overall four-item limit results from a two-item limit within each

hemisphere of the brain, which is also reflected in a distinctive neural signature (the contralateral delay activity, or

CDA) which increases with working memory load but levels off at the two-item limit.

8

existence of qualia), but rather from the special way in which we sometimes think about our own

globally broadcast nonconceptual contents. Specifically, we can refer to our own conscious states in a

way that is completely disconnected from any of our third-person physical, functional, and

representational concepts. But what we thereby refer to, nevertheless, is just globally broadcast

nonconceptual content.

The account of our first-person phenomenal concepts that I now favor follows Prinz (2012) in

thinking of them as a kind of acquaintance-based indexical (Carruthers, 2019). But there is nothing

mysterious about the notion of acquaintance employed here. For it is the very same globally broadcast

nonconceptual contents that can issue in first-order thoughts such as, “That is a lovely shade of red”

that can also issue in higher-order thoughts like, “This could have been reliably caused by the presence

of something green rather than red” or, “There could be a creature that is like me in all respects except

that it lacks this.” In both cases the contents in question are received as input by the conceptual and

planning systems that generate structured thoughts. The only difference is that the first-order thought,

here, depends on tacit deployment of the concept COLOR, whereas the higher-order ones rely on tacit

use of the concept PERCEPTION.6 In both cases globally broadcast nonconceptual contents are made

available to the executive and/or language-production systems that can formulate thoughts, but in the

one case those thoughts are about the contents represented whereas in the other they are about the

representing states themselves.

Although phenomenal concepts, when used in the first person, lack conceptual connections with

physical, functional, and representational concepts, arguably those same concepts can also be given a

distinctive third-personal characterization. Indeed, I did so in the previous paragraph, saying that they

are indexical concepts used to designate one’s own globally broadcast perceptual states through tacit

use of the concept PERCEPTION (or something similar). But why should we suppose that this is so? Indeed,

doesn’t it beg the question against the existence of qualia to say that our phenomenal concepts refer,

not to qualia, but just to globally broadcast nonconceptual content? What reason do believers in qualia

have for accepting such a claim?

One argument builds on considerations of explanatory simplicity. By supposing that phenomenal

concepts are just acquaintance-based indexicals referring to globally broadcast nonconceptual contents

6 Notice that if follows from this that phenomenal concepts aren’t completely disconnected from other mentalistic

concepts. See Carruthers (2017) for discussion of how merely tacit deployment of concepts like PERCEPTION can still

leave us able to think thoughts like, “This [percept of red] might not have had the functional role of perception.”

9

we can explain how the various “hard”-problem thought experiments arise. Yet the explanation in

question just appeals to properties that everyone agrees to exist (global broadcasting, nonconceptual

content, indexical concepts, higher-order concepts). There is simply no need to postulate any additional

properties, such as the existence of non-physical qualia. Indeed, the simplest explanation is that the

explanatory gap is an illusion produced by a disconnect between our third-person physical and

psychological concepts, on the one hand, and our first-personal higher-order indexical ones, on the

other. Put differently, there is an explanatory gap at the level of thought (some thoughts about our

globally broadcast nonconceptual states aren’t entailed by third-personal facts) rather than at the level

of worldly properties. The only real properties that exist are ones that are physical, functional, and

representational. There are no additional first-person-individuated properties.

Here is a somewhat more direct argument for the same conclusion. When using phenomenal

concepts in the first person one can, of course, conceive of the existence of a zombie: someone who is

like oneself in all physical, functional, and representational respects but who nevertheless lacks anything

like this [felt experience]. But given this characterization of a zombie, together with the third-personal

account of phenomenal concepts sketched above, one can see in advance that the zombie, too, will be

able to conceive of a zombie version of itself. Since the higher-order indexicals referring to its own

globally broadcast nonconceptual contents lack connections with its physical, functional, and

representational concepts, the zombie, too, will be able to think, “There could be a creature exactly like

myself in all physical, functional, and representational respects who nevertheless lacks anything like

this” and, “No matter how much I know about the physical, functional, and representational facts

involved in perception, that still won’t explain why those facts give rise to states like this.” Since my

zombie twin, too, will be puzzled by the same “hard”-problem thought experiments as myself, we don’t

need to appeal to anything other than phenomenal concepts in order to explain how those thought

experiments arise.

The success of the phenomenal concept strategy hinges on the existence of phenomenal

concepts, of course. The latter has been denied by both Ball (2009) and Tye (2009), on very similar

grounds. But as shown by Veillet (2012), and as argued in Carruthers (2019), their arguments lose sight

of the fact that phenomenal concepts are grounded in their first-person uses, and that much talk about

phenomenal consciousness doesn’t really involve phenomenal concepts at all. We can, of course,

entertain thoughts about phenomenal states that we have never experienced (think of Mary in her

black-and-white room wondering what it is like to see red), and of course we can form concepts that

defer to the phenomenal concepts of other people. But in doing so we aren’t forming first-person

10

acquaintance-based higher-order indexical concepts, of the sort that give rise to the “hard”-problem

thought experiments.

4. The question of others’ consciousness

Recall the distinction we drew at the outset between the first-personal phenomenal concepts that

define our subject-matter, and the third-person concept we can use to operationalize phenomenal

consciousness for purposes of public discourse. The former are concepts like THIS [introspectively

identified feel of experience]. The latter can be defined as “whatever property explains the hard-

problem thought experiments.” We now know that the latter property is globally broadcast

nonconceptual content (or so I have suggested). So if we ask about the distribution of phenomenal

consciousness across the animal kingdom using the public concept, the answer is straightforward. It is

that animals will possess the property in question (globally broadcast nonconceptual content) to

whatever degree they share the same global workspace architecture as ourselves. And as we will see

later, what exists in nature is a complex cross-cutting multidimensional similarity-space of architectures,

with some creatures sharing some aspects of the global workspace with us to some degree, and others

sharing others. So many other animals turn out to share the publicly-individuated property of

phenomenal consciousness (“whatever property gives rise to the hard problem”) to some significant

degree; and this will, of course, be a matter of degree.

As we emphasized at the outset, however, phenomenal consciousness is basically first-personal.

It is the phenomenal concepts that one can apply introspectively in one’s own case that define the

subject-matter, and that are employed to generate the “hard” problems. And if instead of formulating

the question of animal consciousness using the public operationalized concept, we ask a question like,

“Do other animals undergo states like this [feel of experience]?” then matters are not nearly so

straightforward. Indeed, if the phenomenal concepts that delimit our subject-matter are first-person

acquaintance-based indexicals, then what does it even mean to attribute phenomenal consciousness to

another creature? This question can form our starting point for the remaining discussion.

Suppose one asks oneself, “Do other creatures, too, undergo this type of state?” where the

indexical involved is an acquaintance-based phenomenal concept. What fixes the truth-condition for a

correct answer? Plainly, as Kripke (1980) emphasized, one is not using the indexical concept as a natural-

kind concept. Although that concept does in fact refer to globally broadcast nonconceptual content (if

the considerations adduced in Sections 2 and 3 are correct), the concept is not employed with that

intention. On the contrary, one means just to refer to the subjectively-presented property that one is

11

first-person acquainted with. One doesn’t intend to think through the phenomenally conscious feelings

in question to whatever they really are, to an underlying reality. And one plainly doesn’t intend to refer

to globally broadcast nonconceptual content as such. For one can deploy phenomenal concepts for

one’s own conscious states without even having the concepts of “global broadcasting” or

“nonconceptual content.” On the contrary, one intends to designate just one’s own conscious feelings

themselves.

Here is a related consideration: we can only know that our phenomenal concepts refer to

globally broadcast nonconceptual content if we can first know that other people undergo phenomenally

conscious states. Since the science of consciousness studies presupposes this, our first-person

phenomenal concepts must somehow fix the truth-conditions involved in their projection to other

people in advance of any discovery of what phenomenal consciousness really is.

Phenomenal concepts aren’t natural kind concepts, then. And since they are first-personal,

neither are there any public norms guiding or constraining their use. (Of course we do also have public

concepts for talking about phenomenal consciousness, including the concept PHENOMENAL CONSCIOUSNESS;

but these are secondary to the first-person phenomenal concepts that fix the subject matter in

question.)7 Moreover, since (as it turns out) there are no qualia, the truth-condition of the thought that

another person is undergoing a state of the same sort as this cannot be that the other person’s mind,

too, contains qualia properties. For there are none.

To help us move forward, consider this analogy. While visiting an area of a city to which one

plans to relocate one might exclaim, “So this is the kind of neighborhood where we should live!”

Although the type-indexical “this” here refers to a non-mental property, its use still is, in a sense,

“private.” Plainly the intention is not to pick out a natural kind. Indeed, neighborhoods aren’t the sort of

thing to be natural kinds (even though they may, in part, be composed of them). And nor are there any

public norms guiding or constraining one’s use. Moreover, in the example as I imagine it, one might be

incapable of articulating what it is about the neighborhood in question that seems so right. Now

suppose one asks whether another neighborhood in the city is also of this kind: what fixes the truth-

conditions for a correct answer? I suggest the following: that if the dispositions-to-judge that underlie

one’s initial thought were to be confronted with that other neighborhood in the same background

circumstances of mood, weather, and so on, then they would issue in the same judgment. To project a

private concept from one neighborhood to another requires holding the underlying dispositions-to-

7 See Balog (2009) for the distinction adverted to here between basic and non-basic phenomenal concepts.

12

judge constant, and supposing them to be instantiated in that other neighborhood.

Similarly, then, with the phenomenal concept embedded in the question, “Do others undergo

states of this sort?” Its projection into the mind of another person requires supposing that one’s current

dispositions-to-judge are instantiated in the mind of the other. In fact, the truth condition for, “Agent A

undergoes states of this sort” (where THIS is a phenomenal concept) is as follows: “If the dispositions

underlying my current use of the concept THIS were instantiated in the mind of agent A, then those

dispositions would issue in a positive judgment that agent A has this.”

Where the agent in question is another adult human being, then attributions of states of this

sort will generally be true (modulo factors like blindness and deafness), and one can know them to be

true. This is not just because of the extensive similarities between our respective psychologies,

knowable from a third-person perspective, or the underlying similarities in physical and neural

organization. It is also because those other people, too, can express judgments much like my own. They

can likewise conceive of zombies, and can readily become puzzled by the explanatory gap. This gives me

every reason to think that they, too, employ phenomenal concepts with similar overall properties to my

own. So I can be confident that if the dispositions-to-judge underlying my phenomenal concept THIS

were instantiated in those other minds, then those dispositions would issue in the positive verdict that a

state of this sort is present.

In sum, then, to think that another creature is phenomenally conscious is to think that it

undergoes states like this (where this expresses a first-person phenomenal concept). And that thought

will be true just in case the dispositions-to-judge underlying my use of the concept THIS would issue in a

positive verdict if they were to be instantiated in the mind of the creature in question.

When the creature one is thinking about has a mind significantly unlike one’s own, however,

then attributions of phenomenally conscious states become unevaluable, or truth-valueless. Suppose

one wonders whether Macaque monkeys have phenomenally conscious experiences, for example.

Macaque minds have much in common with our own. Specifically, they have attentional systems that

control the entry of perceptual representations into a form of working memory, just as we do. But there

are also important differences. (One might think of these as differences in the “global” component of

the global workspace.) For their executive function capacities and planning capacities are much weaker

than our own; they lack language altogether; it is unclear whether they engage in System 2 reflective

thinking; and although there is some evidence that they can attribute some forms of mentality to other

creatures, it is controversial whether or not they are capable of thinking about their own mental states

at all (Carruthers & Ritchie, 2012; Carruthers, 2019; Carruthers & Williams, 2019).

13

In supposing, then, that the dispositions-to-judge underlying one’s use of phenomenal concepts

are instantiated in the mind of a Macaque, one might have to suppose that the mind of the Macaque is

quite other than it actually is. For one has to suppose that it is capable of reflectively forming

introspective concepts about its own perceptual states. Maybe Macaques are capable of such a thing.

But suppose I am right that they are not. In that case one cannot evaluate what would happen if

Macaque minds as they actually are were to contain the dispositions-to-judge that underlie my own

phenomenal concepts. If they did possess such dispositions, then their minds would be much more

similar to the human mind than they actually are.

It turns out, then, that projecting first-personal phenomenal consciousness into the mind of

another creature makes an important presupposition. It is presupposed that the mind of the creature in

question could contain the dispositions-to-judge that underlie one’s use of phenomenal concepts

without significantly altering what that mind is like. In connection with most nonhuman animals, that

presupposition is false. And that means that the counterfactuals underlying the attribution of

phenomenally conscious states to those animals can’t be evaluated. So there is no fact of the matter

whether their states are phenomenally conscious or not. Attributions of consciousness to them aren’t

false, but neither true nor false.

It seems that the question of phenomenal consciousness in other creatures turns, in part, on

their capacity for higher-order thoughts. Does this mean that the theory of consciousness defended

here is a version of higher-order thought theory, after all? It does not. Phenomenal concepts are higher-

order, of course. And it is phenomenal concepts that fix the domain of first-personal phenomenal

consciousness. But what those concepts refer to is globally broadcast nonconceptual content. Indeed,

phenomenal consciousness (at least in the human case) is globally broadcast nonconceptual content.

That makes the theory a first-order one. But to ask about phenomenal consciousness in other creatures

(to inquire about the extent of the domain picked out by our first-person phenomenal concepts) we

have to project our phenomenal concepts into their minds. We have to ask whether, if the dispositions

underlying our phenomenal concepts were to be instantiated in those minds, our concepts would apply

to any of their states. It is this that issues in indeterminacy.

5. Marginal cases

I have argued that the truth-conditions of thoughts ascribing phenomenal consciousness to other

creatures involve counterfactuals that project the dispositions-to-judge underlying one’s own

phenomenal concepts into the minds of those creatures. As a result, there are no facts of the matter

14

about phenomenal consciousness in most species of animal. This is because the counterfactuals in

question can’t be evaluated in minds that lack the capacities underlying those dispositions-to-judge.

There will, of course, be marginal and/or otherwise difficult cases, of which I will mention a few

different sorts. In some instances, these will arise because we don’t know enough about the minds of

the creatures in question. We don’t know, for example, the extent to which chimpanzees are capable of

reflective thinking, nor whether they can entertain thoughts about their own current experiences. But a

quite different source of uncertainty is that we might fail to know what capacities are required for a

creature to possess phenomenal concepts like our own. Do such concepts require, not just capacities for

higher-order thought and for reflective thinking, but also a capacity for language? Even if mental-state

concepts aren’t generally language-dependent (as I believe), it could be that phenomenal concepts

themselves are an exception. Perhaps introspectively applied type-indexical thoughts require a natural-

language vehicle.

Other sorts of difficulty arise when one considers the emergence of phenomenal consciousness

in human development from infancy to adulthood. Even if one is quite clear about the cognitive

capacities required for possession of phenomenal concepts, there remains the question of what counts

as a significant change in the nature of the mind under consideration. New-born infants are, of course,

incapable of language, and they are likely incapable of reflective thinking and of higher-order thoughts

about their own mental states. But their minds and brains contain nascent versions of all these

capacities. So when one entertains the counterfactual, “If the dispositions-to-judge underlying my

thought about this were instantiated in the mind of an infant, then they would issue in the judgment

that this is present,” does this require us to suppose that the mind of the infant would be significantly

unlike what it actually is, or not? If I imagine the infant’s mind to contain the capacities underlying those

dispositions, am I imagining a very different kind of mind? It is hard to see what might settle the

question either way.

Yet another sort of potential difficulty arises when we consider the possibility of phenomenal

consciousness in artificial systems. To make the case as vivid as possible, imagine that we have reached

the point where we can build androids that are cognitive isomorphs of adult humans. That is to say, they

comprise systems that are functionally organized in the same manner as the human mind, and are

capable of the same behaviors—including spontaneous expressions of puzzlement about the

explanatory gap and the possibility of zombie versions of themselves. Would this be sufficient for the

truth of the claim that the androids in question are phenomenally conscious? The answer may depend

on one’s views about the metaphysics of dispositions (Choi & Fara, 2018).

15

Recall that to ascribe phenomenal consciousness to another entity requires projecting one’s

own phenomenal concepts into the mind of that entity. It requires us to evaluate a counter-factual

conditional whose antecedent is, “If the dispositions-to-judge underlying my judgment about this were

instantiated in the mind of the other, then … .” And then the question becomes: what has to be true for

one of my cognitive dispositions to be instantiated in the mind of an android? If dispositions are

identical with their categorical bases (Mackie, 1977; Armstrong et al., 1996), then again the counter-

factual becomes unevaluable—at least, provided that the android’s cognitive system is realized in metal

and silicon rather than living neurons. For on this view, my disposition-to-judge, having a categorical

base involving neurons, cannot exist in the android’s mind, unless the android were to be differently

physically constructed. But that requires us to suppose that the android is other than it actually is. So

there is no evaluating whether the android, as it actually is, is phenomenally conscious.

Moreover, even if we retreat to a weaker view of the metaphysics of dispositions, merely

identifying disposition tokens, rather than disposition types, with their categorical bases (Mumford,

1998), the upshot is perhaps not so clear. For that requires us to decide whether one is projecting a

token disposition-to-judge or a type disposition when one supposes that the dispositions underlying

one’s use of phenomenal concepts are instantiated in the mind of the other agent.

Even if everyone were to agree that there is no fact of the matter concerning phenomenal

consciousness in most species of animal, then, there would remain difficult and hard-to-answer

questions at the margins. But are these important questions? Should we even attempt to answer them?

I will now argue for a negative response.

6. Who cares about consciousness?

Recall that on the view that I have been defending, there are no qualia, and there is no mental paint. All

that really exists is globally broadcast nonconceptual content. This can either by thought about as such,

in the third person, or it can be thought about introspectively, employing acquaintance-based indexical

concepts (that is, phenomenal concepts). So nothing magical entered the world with phenomenal

consciousness. Nothing lit up. No spotlight got turned on. Nonconceptual contents just began entering

into some novel functional relationships with other aspects of cognition.

This is why it doesn’t matter that there is no fact of the matter about consciousness in most

species of animal. What changes, as we look across species, are cognitive architectures that are more or

less similar to our own along a number of cross-cutting dimensions (Carruthers, 2019). Creatures can

differ in their executive function abilities, in the forms of memory of which they are capable, in their

16

planning abilities, in the richness of their conceptual repertoire, and much more. Many of these

capacities can vary independently of one another across species. Generalist feeders like bears and

raccoons, for example, will have extensive knowledge of the foodstuffs in their environments and their

likely location and timing. But they may have quite limited planning abilities. Predators that mostly rely

on a single source of prey, in contrast, like wolves and lions, might have quite limited conceptual

knowledge of the environment, but significant planning capacities and strong impulse-inhibition

abilities, enabling them to resist the urge to rush straight towards every potential source of food.

It seems that all mammals and birds, at any rate, have attentional mechanisms, as well as

something resembling human working memory capacities (Carruthers, 2015). Indeed, each of these

things might be significantly more widespread even than this across the animal kingdom. What differs

across species are the mechanisms that control the direction of attention, and the systems that

comprise the global workspace, consuming and responding to the contents that enter working memory.

As one surveys the resulting multidimensional similarity space, capacities get added, subtracted,

enhanced, or reduced. But at no point does any special intrinsically-first-personal property appear. At

some stage (most likely in the primate or hominin lineage) there emerge a combination of abilities for

higher-order thought, reflective thinking, and perhaps language, that together enable the formation of

indexical acquaintance-based concepts about the nonconceptual contents of working memory. And at

that point, projections of one’s own phenomenal concepts into the minds of one’s conspecifics begins to

make sense, and results in true claims about the presence in those creatures of states like this. That is all

there is to it.

Questions about the cognitive capacities of any given species of creature matter, of course, both

for science, and for issues to do with humane treatment as well as animal husbandry (Dawkins, 2017).

But the fact that there is no fact of the matter about phenomenal consciousness (“states like this”) in

most species of animal doesn’t matter for any of those things (Carruthers, 2019). There is no mystery

about the first emergence of special first-person properties, because there are none. And the fact that

there is no fact of the matter isn’t puzzling, either. It results merely from the fact that we cannot

sensibly project our first-person phenomenal concepts into minds that are intrinsically incapable of

forming or deploying them.

Given that facts about phenomenal consciousness don’t matter, then questions about marginal

or difficult cases of phenomenal consciousness don’t matter, either. There might be some philosophical

interest in figuring out whether dispositions are identical to their categorical bases. But this is not

because it matters whether or not consciousness is in principle possible in a non-biological android.

17

Likewise, although there might be some philosophical interest in attempting the clarify the exact

cognitive prerequisites for possession of phenomenal concepts, this isn’t because it matters whether

Macaques or chimpanzees are phenomenally conscious. The question of mentality in artificial systems is

important, as are questions about the cognitive capacities of Macaques and chimpanzees. But the

question of consciousness in these entities isn’t.

I have been arguing that the question whether (and which) other creatures are phenomenally

conscious doesn’t matter, on the grounds that no new first-person property entered the world when

phenomenal consciousness did. Does that mean that I am committed to eliminativism or illusionism

about consciousness (Frankish, 2016)? It does not, or not straightforwardly. Since there are true claims

about phenomenal consciousness, there are facts about phenomenal consciousness. Hence, in one good

sense, phenomenal consciousness is real. When Mary emerges from her black-and-white room, for

example, and thinks, “So this [phenomenally conscious experience] is what it is like to see red,” she

thinks something true. And likewise, I think something true when I think, while morbidly contemplating

my own death, “When I die, all this [phenomenally conscious experience] will disappear.” But the facts

in question are thin rather than thick, tied to and individuated in terms of a particular mode of

conceptualization, while having no reality beyond it.

7. Conclusion

It is time for us to stop thinking about phenomenal consciousness altogether. Certainly it shouldn’t

continue to attract the degree of scientific and philosophical attention it has received up to now. There

are good reasons to believe that phenomenal consciousness (in humans) just is globally broadcast

nonconceptual content, and that the so-called “hard” problem of consciousness arises merely from the

distinctive set of concepts we employ then thinking about it in the first person. Third-person mentality

matters, and understanding the multi-dimensional similarity-space of functionally organized minds

across species matters. But first-personal phenomenal consciousness is just a distraction. So we should

stop caring about it.

References

Armstrong, D.M., Martin, C., & Place, U. (1996). Dispositions: A Debate. London: Routledge.

Baars, B. (1988). A Cognitive Theory of Consciousness. Cambridge, UK: Cambridge University Press.

Ball, D. (2009). There are no phenomenal concepts. Mind, 118, 935-972.

Balog, K. (1999). Conceivability, possibility, and the mind−body problem. The Philosophical Review, 108,

18

497-528.

Balog, K. (2009). Phenomenal concepts. In B. McLaughlin, A. Beckermann, & S. Walter (eds.), The Oxford

Handbook of Philosophy of Mind, Oxford University Press.

Balog, K. (2012). In defense of the phenomenal concept strategy. Philosophy and Phenomenological

Research, 84, 1-23.

Block, N. (2007). Consciousness, accessibility, and the mesh between psychology and neuroscience.

Behavioral and Brain Sciences, 30, 481-499.

Block, N. (2010). Attention and mental paint. Philosophical Issues, 20, 23-63.

Block, N. (2011). Perceptual consciousness overflows cognitive access. Trends in Cognitive Science, 12,

567-575.

Carruthers, P. & Ritchie, J.B. (2012). The emergence of metacognition: affect and uncertainty in animals.

In M. Beran, J. Brandl, J. Perner, & J. Proust (eds.), Foundations of Metacognition, Oxford

University Press.

Carruthers, P. & Veillet, B. (2007). The phenomenal concept strategy. Journal of Consciousness Studies,

14 (9-10), 212-236.

Carruthers, P. & Veillet, B. (2011). The case against cognitive phenomenology. In T. Bayne & M.

Montague (eds.), Cognitive Phenomenology, Oxford University Press.

Carruthers, P. & Veillet, B. (2017). Consciousness operationalized, a debate realigned. Consciousness and

Cognition, 55, 79-90.

Carruthers, P. & Williams, D. (2019). Comparative metacognition. Animal Behavior and Cognition, 6, 278-

288.

Carruthers, P. (2000). Phenomenal Consciousness. Cambridge, UK: Cambridge University Press.

Carruthers, P. (2015). The Centered Mind: What the science of working memory shows us about the

nature of human thought. Oxford, UK: Oxford University Press.

Carruthers, P. (2017). In defense of first-order representationalism. Journal of Consciousness Studies, 24,

5-6, 74-87.

Carruthers, P. (2018). Valence and value. Philosophy and Phenomenological Research, 97, 658-680.

Carruthers, P. (2019). Human and Animal Minds: The consciousness questions laid to rest. Oxford, UK:

Oxford University Press.

Chalmers, D. (1996). The Conscious Mind. Oxford, UK: Oxford University Press.

Choi, S. & Fara, M. (2018). Dispositions. In E. Zalta (ed.), The Stanford Encyclopedia of Philosophy,

<https://plato.stanford.edu/archives/fall2018/entries/dispositions/>.

19

Cohen, M.A. & Dennett, D. (2011). Consciousness cannot be separated from function. Trends in

Cognitive Sciences, 15, 358-364.

Cohen, M.A., Alvarez, G., & Nakayama, K. (2011). Natural-scene perception requires attention.

Psychological Science, 22, 1165-1172.

Cohen, M.A., Dennett, D., & Kanwisher, N. (2016). What is the bandwidth of perceptual experience?

Trends in Cognitive Sciences, 20, 324-335.

Cutter, B. & Tye, M. (2011). Tracking representationalism and the painfulness of pain. Philosophical

Issues, 21, 90-109.

Dawkins, M. (2017). Animal welfare with and without consciousness. Journal of Zoology, 301, 1-10.

Dehaene, S. (2014). Consciousness and the Brain. New York: Viking Press.

Ericsson, A. & Kintsch, W. (1995). Long-term working memory. Psychological Review, 102, 211-245.

Frankish, K. (2016). Illusionism as a theory of consciousness. Journal of Consciousness Studies, 23 (11-

12), 11-39.

Kapoor, V., Dwarakanath, A., Safavi, S., Werner, J., Besserve, M., Panagiotaropoulos, T., & Logothetis, N.

(2020). Decoding the contents of consciousness from prefrontal ensembles. BioRxiv preprint.

Kripke, S. (1980). Naming and Necessity. Oxford, UK: Blackwell.

Landman, R., Spekreijse, H., & Lamme, V. (2003). Large capacity storage of integrated objects before

change blindness. Vision Research, 43, 149-164.

Lau, H. & Rosenthal, D. (2011). Empirical support for higher-order theories of conscious awareness.

Trends in Cognitive Sciences, 15, 365-373.

Lewis-Peacock, J., Drysdale, A., Oberauer, K., & Postle, B. (2012). Neural evidence for a distinction

between short-term memory and the focus of attention. Journal of Cognitive Neuroscience, 24,

61-79.

Loar, B. (1990). Phenomenal states. In J. Tomberlin (ed.), Philosophical Perspectives: Action theory and

philosophy of mind, Atascadero, CA: Ridgeview Press.

Mackie, J.L. (1977). Dispositions, grounds and causes. Synthese, 34, 361-370.

Mashour, G., Roelfsema, P., Changeux, J-P., & Dehaene, S. (2020). Conscious processing and the global

neuronal workspace hypothesis. Neuron, 105, 776-798.

Michel, M. (2017). Methodological artefacts in consciousness science. Journal of Consciousness Studies,

24 (11-12), 94-117.

Mumford, S. (1998). Dispositions. Oxford: Oxford University Press.

Panagiotaropoulos, T., Deco, G., Kapoor, V., & Logohetis, N. (2012). Neuronal discharges and gamma

20

oscillations explicitly reflect visual consciousness in the lateral prefrontal cortex. Neuron, 74,

924-935.

Papineau, D. (2002). Thinking about Consciousness. Oxford, UK: Oxford University Press.

Prinz, J. (2012). The Conscious Brain. Oxford, UK: Oxford University Press.

Rosenthal, D. (2005). Consciousness and Mind. Oxford, UK: Oxford University Press.

Salti, M., Monto, S., Charles, L., King, J-R., Parkkonen, L., & Dehaene, S. (2015). Distinct cortical codes

and temporal dynamics for conscious and unconscious percepts. eLife, 4, e05652.

Sergent, C., Wyart, V., Babo-Rebelo, M., Cohen, L., Naccache, L., & Tallon-Baudry, C. (2013). Cueing

attention after the stimulus is gone can retrospectively trigger conscious perception. Current

Biology, 23, 150-155.

Sligte, I., Scholte, H.S., & Lamme, V. (2008). Are there multiple visual short-term memory stores? PLoS

ONE, 3, e1699.

Sperling, G. (1960). The information available in brief visual presentations. Psychological Monographs:

General and Applied, 74, 1-29.

Thibault, L., van den Berg, R., Cavanagh, P., & Sergent, C. (2016). Retrospective attention gates discrete

conscious access to past sensory stimuli. PLoS One, 11, e0148504.

Tononi, G. & Koch, C. (2015). Consciousness: Here, there, and everywhere? Philosophical Transactions of

the Royal Society B, 370, 20140167.

Tononi, G. (2008). Consciousness as integrated information: A provisional manifesto. Biological Bulletin,

215, 216-242.

Trübutschek, D., Marti, S., Ojeda, A., King, J-R., Mi, Y., Tsodyks, M., & Dehaene, S. (2018). A theory of

working memory without consciousness or sustained activity. eLife, 6, e23871.

Tsubomi, H., Fukuda, K., Watanabe, K., & Vogel, E. (2013). Neural limits to representing objects still

within view. Journal of Neuroscience, 33, 8257-8263.

Tye, M. (1995). Ten Problems of Consciousness. MIT Press.

Tye, M. (2000). Consciousness, Color, and Content. MIT Press.

Tye, M. (2009). Consciousness Revisited: Materialism without phenomenal concepts. MIT Press.

Veillet, B. (2012). In defense of phenomenal concepts. Philosophical Papers, 41, 97-127.

Wu, C-C. & Wolfe, J. (2018). A new multiple object awareness paradigm shows that imperfect

knowledge of object location is still knowledge. Current Biology, 28, 3430-3434.

Xia, Y., Morimoto, Y., & Noguchi, Y. (2016). Retrospective triggering of conscious perception by an

interstimulus interaction. Journal of Vision, 16, 3.

21